Daniel's question this week starts from a very specific place. You see the headlines, Ukraine, Gaza, Sudan, and it feels like the world is on fire. He wants to know if it actually is more on fire than usual, and who gets to decide what counts as a fire in the first place. He asked whether there's somebody at the UN keeping a spreadsheet, what the formal definition of a war even is, whether we only count international wars between sovereign states, how we handle countries that are involved in reality but not officially, whether formal declarations of war still happen, and what the long-term trend actually shows versus the perception of unprecedented conflict.
That last part is the tension. The perception is that everything is collapsing into violence, and the data says something much stranger.
So today we're looking at who actually counts wars, what definitions they use, and what the numbers really say. Because the answer is more complicated than you'd think, and the complication starts with the word war itself.
Right, because war is not a technical term. It's a legal term, a political term, a colloquial term. If you ask five different organizations how many wars are happening right now, you get five different numbers, and they're all defensible.
The UN doesn't keep a spreadsheet. That's the first thing to clear up.
No, the UN responds to crises ad hoc. The Security Council meets, it passes resolutions, it authorizes peacekeeping missions. But nobody at the UN is sitting in a basement updating a master list of active wars. The systematic tracking is done by academic institutions, and the big one is the Uppsala Conflict Data Program, UCDP, out of Sweden. They've been doing this since the nineteen eighties.
And they're the ones whose numbers usually end up in the news, even if the news doesn't name them.
UCDP has a very specific definition. They count what they call armed conflict, which requires at least twenty five battle-related deaths in a calendar year, and at least one of the parties has to be a state government. So right away, they're excluding purely non-state violence. The cartel wars in Mexico, gang violence in Central America, those don't count because no state government is a party.
So twenty five deaths is the floor. What's the threshold for actually calling something a war?
UCDP reserves the term war for conflicts that hit a thousand battle-related deaths in a single year. So anything between twenty five and nine hundred ninety nine is an armed conflict. Once you cross a thousand, it's a war. And these thresholds are arbitrary in the sense that someone had to pick a number, but they're not random. Twenty five is low enough to catch things early, and a thousand is a conventional threshold that's been used in conflict research for decades.
But the arbitrariness matters. A conflict with twenty four deaths doesn't exist in their data. A conflict with nine hundred ninety nine deaths is not a war. The number shapes the story.
It does, and the people who run UCDP are completely upfront about that. They know the threshold problem. The point is consistency over time, not metaphysical precision. If you use the same definition for forty years, you can compare trends even if the definition itself is imperfect.
Daniel also asked about the categories. State versus non-state, and something about one-sided violence.
UCDP tracks three types of violence. State-based conflict is what we've been talking about, a state government fighting another state or an organized armed group. Non-state conflict is between two organized groups, neither of which is a state, think rebel factions fighting each other. And one-sided violence is a state or an armed group attacking unarmed civilians. The numbers look completely different depending on which type you're counting. If you include all three, the world looks much more violent than if you only count state-based wars.
And most people, when they say how many wars are going on, they're picturing state-based conflicts. Countries fighting countries, or countries fighting rebels.
Right, and that's already a narrow slice. But even within state-based conflict, there's the involvement problem Daniel raised. Countries that are involved in reality but not officially.
The US in Yemen is the classic example.
The US has been providing intelligence, refueling, and munitions to the Saudi-led coalition in Yemen for years. But the US is not formally a party to that conflict. UCDP handles this by coding external support and secondary parties. So the US gets a code indicating it's providing support, but the conflict itself is coded as the Yemeni government against the Houthis, with the Saudi coalition as a secondary party. The US is in the data, but it's buried in the fine print.
Which means if you do a simple count of how many wars the US is in, you get a very misleading number.
You get zero, technically. The US hasn't formally declared war since World War Two. But the US has been involved in dozens of conflicts since then. Korea, Vietnam, Iraq, Afghanistan, Libya, Syria, Yemen. None of them involved a declaration of war.
And that's the declaration question. Daniel asked whether countries still do that, or if we just start attacking and then retrospectively construct the declaration.
The last formal declaration of war by the United States was in nineteen forty two, against Romania, Bulgaria, and Hungary, if I remember correctly. Since then, countries have largely stopped declaring war. The legal mechanism still exists, it's in the UN Charter, but it's almost never used.
Russia didn't declare war on Ukraine.
Russia called it a special military operation. That was the official framing from the Kremlin. But UCDP codes the Russia-Ukraine conflict as a state-based armed conflict starting in twenty fourteen, with the annexation of Crimea and the fighting in Donbas, and then escalating to war-level intensity in twenty twenty two. The coding is based on the reality of fighting and the death toll, not on what the Russian government calls it.
So the legal label is almost irrelevant to the data. What matters is the bodies.
The bodies, the parties, and the organization. UCDP requires that at least one party be a state government, and that the other party be an organized armed group with a name, a command structure, and the ability to sustain military operations. That last part is important. It excludes riots, spontaneous uprisings, and disorganized violence.
So if a mob storms a government building and twenty six people die, that's not necessarily an armed conflict.
Correct. It has to be sustained, organized, and meet the death threshold. The organization requirement is actually one of the harder things to code. How do you determine if a group has a command structure? You look at whether they issue statements, whether they have a recognizable leadership, whether they control territory, whether they negotiate. It's judgment calls all the way down.
And that's where the human element comes in. Somebody is reading news reports and coding these things.
Somebody in Uppsala, or one of their partner institutions, is reading news wires, NGO reports, government statements, and making a decision about whether a particular group qualifies as an organized armed group. They have a coding manual, it's hundreds of pages long, but at the end of the day, a human being is applying judgment.
Now let's get to the numbers. What does the long-term data actually show?
UCDP's time series goes back to nineteen forty six. The number of active state-based armed conflicts peaked in the early nineteen nineties, right after the Soviet Union collapsed. There were over fifty active conflicts in some of those years. Then it declined through the two thousands, bottomed out around thirty or so, and has been rising again in the twenty twenties.
And the most recent data point?
Twenty twenty three recorded fifty nine active state-based conflicts. That's the highest number since the Cold War. But here's the key thing, it's still below the early nineties peak in terms of the raw count, and more importantly, the deadliness of conflict today is nowhere near what it was in the mid twentieth century.
So the world is not unprecedentedly at war. It's at a level comparable to the early two thousands, or the early nineties, depending on how you slice it.
Right. The perception of unprecedented chaos is driven by the fact that several high-profile conflicts are happening simultaneously in places that get a lot of media coverage. Ukraine, Gaza, Sudan. But the statistical reality is that we've been here before, and we've been worse.
And then there's the other major data source, the Heidelberg Institute.
HIIK, the Heidelberg Institute for International Conflict Research. They use a much broader definition. They count political conflicts of all intensities, including non-violent disputes, and they don't require a state to be a party. Their twenty twenty four report counted something like two hundred twenty six conflicts globally. That's a huge number compared to UCDP's fifty nine.
But they're not counting wars. They're counting conflicts.
HIIK's definition includes things like diplomatic standoffs, trade disputes that escalate, domestic political crises. Only a fraction of their two hundred twenty six would meet UCDP's threshold for armed conflict, let alone war. So when you see a headline saying there are over two hundred conflicts in the world, that's HIIK's number, and it's measuring something different from what most people mean by war.
And the trend lines point in different directions. UCDP shows a decline from the Cold War peak and a recent rise. HIIK shows a steady increase since nineteen forty five.
Because HIIK is counting more and more types of things as the world becomes more interconnected and more disputes get documented. It's not that the world is necessarily more violent, it's that the net is wider. More things get caught.
This is the Long Peace debate. Steven Pinker's argument that violence has declined over centuries.
Pinker's thesis is contested, but the post-nineteen forty five period did see fewer inter-state wars than the early twentieth century. The two world wars were anomalies in their scale. Since then, most conflicts have been civil wars, not wars between sovereign states. The recent exceptions, Russia-Ukraine, Israel-Hamas, those are high-profile because inter-state war had become rare.
And that rarity is part of why the perception is so strong. When something rare happens, it gets more attention.
The new wars thesis captures this. Modern conflicts are more likely to be civil wars with external involvement, proxy dynamics, non-state actors. They're messier, they last longer, and they're harder to resolve. The average civil war since nineteen forty five has lasted something like ten years. Inter-state wars, when they happen, are often shorter but more intense.
The proxy dynamic is what Daniel was getting at with the involvement question. Iran in Syria, the US in Yemen, Russia in Libya. These countries are fighting without being officially at war.
And the coding struggles to capture that. UCDP has a category for internationalized civil wars, which is a civil war where one or both sides receive external support from a foreign state. But the intensity of that support varies enormously, from a few advisors to full air campaigns. The data flattens all of that into a single code.
So the answer to Daniel's question about how many countries are currently at war depends entirely on how you define at war. If you mean formally declared war, the answer is almost zero. If you mean involved in armed conflict, the answer is dozens.
And if you mean involved in someone else's conflict through support, the answer is probably over a hundred. The United States alone has military personnel in something like eighty countries, though most of that is basing and training, not combat.
The UN's role in all this is worth naming. The Security Council doesn't maintain a conflict database. The peacekeeping department has operational data, how many peacekeepers are deployed where, but that's not the same as tracking wars.
The real tracking is done by UCDP and PRIO, the Peace Research Institute Oslo. They're academic institutions funded by research grants. The UN actually relies on their data when it needs conflict statistics. The UN's own reports often cite UCDP numbers.
So the most authoritative count of global wars comes from a university department in Sweden.
And a research institute in Norway. The Scandinavians have cornered the market on counting violence. It's a very specific niche.
There's a joke in there somewhere about peace research being a growth industry.
It's been growing since nineteen forty five, according to HIIK.
Let's talk about the gap between the data and the headlines. Daniel's perception is that the world is unprecedentedly at war. The data says the number of active conflicts is high but not unprecedented. What's driving the gap?
Several things. First, the media environment. We now have real-time coverage of conflicts that would have been invisible thirty years ago. Satellite imagery, social media, citizen journalism. A massacre in a remote part of Sudan is now on your phone within hours. That creates a sense of omnipresent violence.
Second, the specific conflicts that are happening right now are in places that matter to Western audiences.
Ukraine is the largest land war in Europe since nineteen forty five. That's historic. Gaza is a conflict involving a state that gets enormous media attention. Sudan is a humanitarian catastrophe of enormous scale. These aren't peripheral conflicts. They're central to the news agenda.
Third, there's the recency bias. People forget that the early two thousands had Afghanistan, Iraq, Darfur, the Congo wars, Sri Lanka, Colombia, and a dozen other conflicts happening simultaneously.
The Congo wars alone killed over five million people. That's the deadliest conflict since World War Two, and it barely registered in Western media compared to Ukraine. The perception of conflict is not evenly distributed.
And fourth, there's the intensity question. The number of conflicts matters, but so does the death toll. And the death toll from state-based conflict today is lower than it was during the Cold War, even though the number of conflicts is high.
The absolute peak in battle deaths since nineteen forty six was in the late nineteen forties and early nineteen fifties, during the Chinese Civil War and the Korean War. Then there were spikes in the late sixties and early seventies, Vietnam, the Bangladesh Liberation War. The early nineties had a spike with the Rwandan genocide and the Yugoslav wars. Today's death toll is high compared to the two thousands, but it's not historically unprecedented.
So the world feels like it's on fire because we're seeing the fires more clearly, not because there are more fires.
That's the core insight. The data doesn't support the claim that we're in an unprecedented era of global conflict. What it supports is that we're in an era of unprecedented visibility of conflict.
And the visibility itself has political consequences. The perception of unprecedented conflict drives policy responses.
It drives defense budgets, it drives intervention decisions, it drives refugee policy. If policymakers believe the world is uniquely dangerous, they act accordingly. That's why getting the numbers right matters. The numbers are used to justify real decisions with real consequences.
Daniel asked about the mechanism of declaration. Do countries still do it?
Almost never. The last formal declaration of war by any major power was, I think, the Soviet Union declaring war on Japan in nineteen forty five, right at the end of World War Two. Since then, countries have used other legal frameworks. Authorizations for the use of military force, like the US Congress did after nine eleven. UN Security Council resolutions. Self-defense claims under Article fifty one of the UN Charter.
The UN Charter actually made declarations of war legally awkward.
The Charter prohibits the use of force except in self-defense or with Security Council authorization. So declaring war is almost an admission that you're violating the Charter. Countries prefer to frame their actions as self-defense or as enforcement of international law. The legal label matters for legitimacy, but it's disconnected from the reality of fighting.
Russia's special military operation is the perfect example. The entire framing was designed to avoid the legal implications of calling it a war. And yet UCDP codes it as a war because the death toll and the parties meet the threshold.
The data doesn't care what you call it. That's both the strength and the limitation of the UCDP approach. It's objective in the sense that it ignores legal labels, but it's still subjective in the application of thresholds and the coding of parties.
Let's get concrete about the current numbers. UCDP recorded fifty nine active state-based conflicts in twenty twenty three. What does that actually include?
It includes everything from the Russia-Ukraine war to the civil war in Myanmar to the fighting in the Sahel to the conflict in Colombia between the government and various armed groups. The list is long and heterogeneous. Some of these conflicts are massive, with tens of thousands of deaths. Others are small, with just over the twenty five death threshold.
And the twenty twenty four data?
Not fully published yet, but the trend is upward. The conflict in Gaza, the escalation in the Middle East, the ongoing war in Sudan, the continued fighting in Ukraine. Twenty twenty four and twenty twenty five are likely to show even higher numbers than twenty twenty three.
So the recent rise is real. The question is whether it's a spike or a new normal.
That's the open question in the field. Some researchers argue we're seeing a structural shift, driven by great power competition, climate stress, and the proliferation of armed groups. Others argue it's a cyclical uptick within the normal range of variation. The honest answer is we don't know yet.
And the definitional mess makes it harder to answer. If you broaden the definition to include non-state conflicts and one-sided violence, the trend looks different than if you only count state-based wars.
UCDP's data shows that non-state conflicts have been rising faster than state-based conflicts. The number of armed groups fighting each other, without a state involved, has increased significantly since the two thousands. That's a different kind of violence, and it's often invisible in the headlines.
Because there's no state to blame, no government to hold accountable.
And no clear narrative. A war between two rebel factions in the Central African Republic doesn't fit the template of good versus evil that drives media coverage. It's just violence.
The one-sided violence category is even more disturbing. That's states or armed groups attacking civilians.
Genocide, ethnic cleansing, massacres. UCDP tracks these separately because the dynamics are different from battle deaths. The Rwandan genocide is the canonical example. Most of the victims were not combatants. They were civilians killed by the state or by militias.
And if you include one-sided violence in your count of global conflict, the picture changes dramatically.
The deadliest episodes of the last thirty years have often been one-sided violence rather than conventional war. Rwanda, Srebrenica, Darfur, the Rohingya. These are not wars in the traditional sense, but they're the worst things that have happened.
So when Daniel asks how many wars are going on, the answer depends on whether he's asking about wars, armed conflicts, or mass atrocities.
And the honest answer is that all three are happening simultaneously, at different scales, in different places, and the data captures them imperfectly.
Let's talk about the UN's actual role. Daniel asked if there's somebody at the UN keeping a spreadsheet. The answer is no, but the UN does have data.
The UN Department of Peace Operations tracks peacekeeping missions. The Office for the Coordination of Humanitarian Affairs tracks humanitarian crises. The Security Council tracks the situations on its agenda. But none of these is a comprehensive conflict database. The UN's own reports on global conflict trends cite UCDP and PRIO.
The UN outsources its conflict counting to Scandinavian academics.
Which is actually a good thing. The academics have independence, methodological consistency, and a long time series. The UN is a political body. Any database it maintained would be subject to political pressure. Member states would object to being listed as parties to conflicts. The academic institutions are insulated from that.
Though not entirely. UCDP's coding decisions are sometimes contested. Countries have objected to being listed as supporting certain conflicts.
The researchers have to make judgment calls about contested facts. Was a particular group an organized armed group at a particular time? Was a particular death a battle death or a civilian casualty? These are hard calls, and reasonable people can disagree.
The point is that the numbers are not objective facts. They're human choices, applied consistently over time.
That's the key insight. The numbers are constructed. They're constructed using explicit rules and transparent methodology, but they're constructed nonetheless. The twenty five death threshold, the thousand death war threshold, the state party requirement, all of these are choices.
Different choices produce different numbers. HIIK's two hundred twenty six conflicts versus UCDP's fifty nine. Both are defensible. They're just measuring different things.
The public discourse rarely makes that distinction. A headline says there are over two hundred conflicts in the world, and people think that means two hundred wars. It doesn't.
The perception gap is partly a translation problem. The data is translated into headlines, and the nuance gets lost.
Partly a selection problem. The conflicts that get covered are the ones that matter to Western audiences. The conflicts that don't get covered, the ones in the Central African Republic or Papua New Guinea, they're still in the data, but they're invisible to the public.
The public perception of global conflict is shaped by media coverage, which is shaped by geopolitical salience, which is shaped by factors that have nothing to do with the actual level of violence.
The visibility of conflict is a function of media attention, not of death tolls. The Congo wars killed millions and got a fraction of the coverage that Ukraine gets. That's not a criticism of the coverage of Ukraine, it's just a fact about how attention works.
The attention itself feeds back into the perception. The more coverage, the more people think the world is at war. The more people think the world is at war, the more they demand coverage. It's a feedback loop.
That's the perception trap Daniel was pointing at. The headlines shape the perception, and the perception shapes the policy, and the policy shapes the headlines.
Let's step back and think about what this means for understanding the world. If the numbers are constructed, and the perception is distorted, what can we actually know?
We can know the direction of travel. The number of state-based conflicts is rising. The number of non-state conflicts is rising faster. The death toll is up from the two thousands but down from the Cold War peaks. The nature of conflict is shifting from inter-state wars to internal conflicts with external involvement.
Those are real trends, even if the exact numbers are fuzzy.
They're actionable trends. If you're a policymaker, you care about the direction, not the exact count. You care that conflicts are becoming more internal, more prolonged, more internationalized. That tells you something about what kind of military and diplomatic tools you need.
The exact count matters less than the shape of the curve.
The shape of the curve, and the composition. What kinds of conflicts, involving what kinds of actors, in what regions. That's where the analytical value is.
Daniel's question about declarations also points to something deeper. The legal framework for war is obsolete, but we haven't replaced it.
The UN Charter was written in nineteen forty five, for a world of sovereign states fighting sovereign states. The world we actually live in is one of non-state actors, proxy wars, and gray zone conflicts. The legal categories don't fit the operational reality.
The data categories are struggling to keep up too. UCDP has had to add new categories over the years, internationalized civil wars, external support, to capture the reality of modern conflict.
The categories are always lagging behind the reality. By the time you've defined a new type of conflict, the world has moved on to a new type.
Which is why the whole enterprise is inherently imperfect. You're trying to impose order on chaos, and the chaos keeps evolving.
But the alternative, not counting at all, is worse. An imperfect count is better than no count. At least you can see the shape of the thing, even if the edges are blurry.
That's the defense of the whole enterprise. Imperfect measurement is better than no measurement.
The imperfections are known. The researchers at UCDP are the first to tell you about the threshold problem, the coding challenges, the judgment calls. They're not pretending to have objective truth. They're providing a consistent, transparent, defensible approximation.
Which is more than the headlines provide.
Far more.
Hilbert: The database was bought by a defense contractor.
What?
Hilbert: Nineteen ninety eight. I was doing data entry for a peace studies NGO in Boston. We were trying to compile a complete list of global conflicts. Not wars, conflicts. Anything that looked like it might turn into something. The boss had a theory that if you caught conflicts early enough, you could predict which ones would escalate. So we coded everything. Border skirmishes, ethnic tensions, trade disputes, water rights. Six months of reading news wires and filling out forms.
How many entries did you end up with?
Hilbert: Over three hundred. And the definitions kept changing. One week a border skirmish counted, the next week it didn't. The boss would read a new article and decide we needed a new category. Low-intensity conflict, pre-conflict, conflict-prone. By the end, the database was useless. Nobody could agree on what anything meant.
Then it got sold.
Hilbert: To a defense contractor. They wanted the list of conflict-prone regions. For market analysis, they said. A few years later, I saw a news report about arms sales to a country that was on our list. Listed as at risk. The contractor had used our database to justify the sale. Preventive security assistance, they called it.
The numbers were used to sell weapons to a country that was flagged as potentially unstable.
Hilbert: That's the thing about conflict statistics. They're not neutral. Somebody uses them. Somebody decides what to count, and somebody else decides what to do with the count. The counting itself is the easy part. The hard part is knowing who's going to read the spreadsheet.
What they're going to do with it.
Hilbert: I quit after that. Went back to driving a delivery van. At least with a van, you know where you're going.
The data doesn't tell you what to do with it. It just sits there until somebody picks it up.
Hilbert: The people who pick it up usually have a reason. A budget request, a sales pitch, a policy paper. The numbers get bent to fit the purpose. Not because the numbers are wrong, but because the purpose comes first.
The definitional mess is not just an academic problem. It's a political problem.
Hilbert: It's a money problem. The definition determines what gets funded. Count something as a war, and the funding follows. Count it as a low-intensity conflict, and it gets ignored. The people doing the counting know this. They're not naive.
That's the cynical layer. The numbers are constructed, and the construction has consequences.
Hilbert: The NGO I worked for thought they were doing peace research. What they were actually doing was building a targeting list. It just took a few years for the targeting to start.
The people who built it had no idea.
Hilbert: The boss knew. He just didn't care. He wanted the grant renewed. The contractor wanted the data. The country wanted the weapons. Everyone got what they wanted except the people in the country.
That's the thing about the perception gap. It's not just about media coverage. It's about who benefits from the perception.
Hilbert: The weapons industry benefits from the perception that the world is at war. The peace researchers benefit from the perception that they're measuring something real. The politicians benefit from the perception that they're doing something about it. The only people who don't benefit are the ones getting shot.
When we ask how many wars are going on, we should also ask who wants to know, and why.
Hilbert: That's the question. Not just what counts, but who's counting, and what they're going to do with the number.
The count is never just a count. It's always an argument.
Hilbert: It's always a sales pitch. Sometimes for peace, sometimes for war. Usually for money.
The spreadsheet just sits there, waiting for someone to use it.
Hilbert: The spreadsheet I built is probably still in a database somewhere. Still flagged. Still at risk. Still generating sales.
That's the thing about data. It outlives its creators.
Hilbert: And its intentions.
The one thing I'd take from this is that the next time you see a headline about the world being at war, remember that the number behind it is a human choice. Somebody decided what to count, somebody decided what threshold to use, somebody decided which conflicts mattered enough to include. The number is not a fact. It's an argument. And the argument is always being made by someone, for a reason.
The reason is rarely just to tell you how many wars there are.
The open question for me is whether better data will solve the perception problem or just create new ones. Satellites are getting better, AI can process conflict data faster, we might get real-time conflict tracking that bypasses the definitional debates. But then somebody has to decide what the algorithm counts. The definitions don't go away, they just get embedded in code.
That's the future I'd watch. The counting will get more automated, but the choices will still be human. Somebody will still have to decide what counts as a war. And somebody will still have to decide what to do with the number.
Thanks to our producer, Hilbert Flumingtop. This has been My Weird Prompts. If you enjoyed the show, email us at show at my weird prompts dot com. We'll be back soon.
